3178
Biochemistry 1998, 37, 3178-3186
Mobility of TrkA Is Regulated by Phosphorylation and Interactions with the Low-Affinity NGF Receptor† David E. Wolf,*,‡ Christine McKinnon-Thompson,‡ Marie-Claire Daou,‡ Robert M. Stephens,§ David R. Kaplan,| and Alonzo H. Ross‡ Departments of Physiology and Pharmacology, UniVersity of Massachusetts Medical School, 55 Lake AVenue North, Worcester, Massachusetts 01655, ABL-Basic Research Program, National Cancer Institute, Frederick Cancer Research and DeVelopment Center, Frederick, Maryland 21701, and Montreal Neurological Institute, Montreal, Quebec H3A 2B4, Canada ReceiVed August 5, 1997; ReVised Manuscript ReceiVed NoVember 25, 1997
ABSTRACT: The nerve growth factor (NGF) receptor is a complex of two proteins, gp75 and the tyrosine kinase TrkA. Using fluorescence recovery after photobleaching, we have studied the diffusion properties of the TrkA receptor. For PC12 cells that express both gp75 and TrkA, TrkA was relatively immobile with only 28 ( 1% of receptor molecules free to diffuse with D ) (3.64 ( 0.23) × 10-9 cm2/s. Addition of NGF decreased the mobile fraction to 21 ( 1% with D ) (4.11 ( 0.18) × 10-9 cm2/s. Using the Sf9 baculovirus expression system, we were able to study TrkA in the absence and presence of gp75. On Sf9 cells, TrkA showed a mobile fraction of 46 ( 2% with D ) (2.64 ( 0.21) × 10-9 cm2/s in the absence of gp75 and 43 ( 2% with D ) (2.31 ( 0.25) × 10-9 cm2/s in its presence. Thus, gp75 did not alter TrkA mobility. Addition of NGF to the medium approximately halved the mobile fraction for TrkA in both the absence and presence of gp75. However, using a kinase-deficient mutant of TrkA, we found that ligand-induced immobilization requires an active kinase in the absence of gp75 but not in its presence. In addition, using point mutations at specific TrkA autophosphorylation sites, we determined that mobility is controlled by multiple phosphorylation sites, but the SHC binding site at Y490 may be particularly important for ligand-induced immobilization of TrkA. Therefore, two mechanisms lead to NGF-induced immobilization of TrkAsthe first resulting from autophosphorylation of TrkA and the second occurring through TrkA’s association with gp75.
Ligand-induced immobilization of receptor tyrosine kinases during receptor-mediated activation was originally described for the nerve growth factor (NGF)1 (1), epidermal growth factor (2, 3), and insulin receptors (2, 3). For each of these receptors, binding of the corresponding growth factor induced tyrosine activation and immobilization of the receptor. Although there was considerable evidence for this phenomenon, the mechanism of immobilization was unknown. It was not clear whether immobilization resulted from receptor aggregation, phosphorylation, or association with other membrane components. In recent years, our understanding of receptor activation and signal tranduction has grown immensely, but the topic of ligand-induced immobilization of receptors has remained a mystery. In this † This work was supported in part by National Institutes of Health Grants NS28760 (to D.E.W. and A.H.R.) and NS21716 (A.H.R.). * To whom all correspondence should be addressed: Department of Physiology, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655. (tel) 508-842-8921; (fax) 508842-9632; (e-mail)
[email protected]. ‡ University of Massachusetts Medical School. § National Cancer Institute. | Montreal Neurological Institute. 1 Abbreviations: NGF, nerve growth factor; BDNF, brain-derived neurotrophic factor; D, diffusion coefficient; DMEM, Dulbecco’s Minimum Essential Medium; FRAP, fluorescence recovery after photobleaching; mAb, monoclonal antibody; moi, multiplicity of infection; na, numerical aperature; PLCγ1, phospholipase Cγ1; PKC, protein kinase C; R, recovery (in percent).
paper we examine immobilization of nerve growth factor receptor TrkA in relation to ligand binding, homo- and heterodimerization, and receptor phosphorylation. NGF is an essential factor required for the development, maintenance, and repair of the nervous system (4). All of the actions of NGF are thought to result from the action of two cell surface receptors, gp75 and TrkA (5, 6). Gp75 is a 75 000-Da glycoprotein with a cysteine-rich extracellular domain, a single transmembrane domain, and a 155-amino acid cytoplasmic domain (7, 8). There is evidence that gp75 can activate signal transduction through cytoplasmic kinases (9, 10), sphingomyelin hydrolysis (11), and transcription factor NF-κB (12, 13). The second NGF receptor is TrkA, a 140 000-Da glycoprotein with a single transmembrane domain and a 304-amino acid cytoplasmic domain that contains a tyrosine kinase domain (6). NGF activates TrkA autophosphorylation but does not result in tyrosine phosphorylation of gp75 (14, 15). Tyrosine autophosphorylation sites on TrkA include the following: the SHC binding domain at Y490, the phospholipaseCγ1 (PLCγ1) binding domain at Y785, and three sites within the kinase domain Y670, Y674, and Y675 that regulate phosphorylation activity (16, 17). A number of experiments indicate that the two NGF receptors interact and synergize with each other. Transgenic mice that lack gp75 have neurological deficiencies involving the death of TrkA-positive sensory neurons (18). Both gp75
S0006-2960(97)01925-9 CCC: $15.00 © 1998 American Chemical Society Published on Web 02/12/1998
NGF Receptor TrkA Membrane Diffusion and TrkA undergo retrograde transport from the synapse to the cell body (19-21). In cell culture models, coexpression of gp75 enhances the biological response of cells expressing TrkA (22-26). Although Jing et al. (27) reported that TrkA is sufficient for high-affinity NGF binding, Hempstead et al. (28) found that expression of both gp75 and TrkA is required for high-affinity binding. In addition, coexpression of gp75 with TrkA enhances the rate of NGF binding (29). The antibody to gp75 inhibits responses of PC12 cells to NGF (30, 31). TrkA is essential for NGF activation of protein kinases associated with the cytoplasmic domain of gp75 (9, 10). The synergistic action of these receptors has led to the hypothesis that gp75 and TrkA form a heteromolecular complex which is the high-affinity NGF receptor (28). Using fluorescence recovery after photobleaching (FRAP), we have shown that gp75 in coexpressing cells is immobilized through interactions with TrkA (32). This immobilization, like high-affinity binding (28), does not occur for gp75(Xba), a mutant form of gp75 lacking most of its intracellular domain (22). NGF causes some immobilization of gp75 probably due to cross-linking between receptors by dimeric NGF in either the absence or presence of TrkA (32). In addition, NGF increases the diffusion rate of gp75 and gp75(Xba) in the presence of TrkA. These results suggest that the requirements for gp75-TrkA complex formation are different from those for immobilization. Extracellular and/ or transmembrane domains of gp75 and TrkA are critical for interaction and complex formation. However, receptor immobilization requires, in addition, intact intracellular domains. By a copatching assay, we have demonstrated that gp75 and TrkA interact with one another to form a heteromolecular complex. In contrast, gp75 does not complex with the brainderived neurotrophic factor receptor, TrkB, the plateletderived growth factor, PDGF, receptor, or the Drosophila receptor, Torso (33). Complexing does not require the kinase activity of TrkA. Using chimeric TrkA receptors (33), we have shown that complexing between gp75 and TrkA is most influenced by the extracellular domains. This paper focuses on the mechanism by which NGF leads to immobilization of TrkA. We show that NGF binding to TrkA leads to receptor immobilization by a kinase-dependent mechanism. TrkA receptors modified at individual autophosphorylation sites show altered states of immobilization, indicating that autophosphorylation modulates lateral mobility of TrkA. For cells coexpressing gp75 and TrkA, NGF also leads to immobilization of TrkA, but autophosphorylation of TrkA is not required. These data suggest two underlying causes for receptor immobilization. The first is dependent on autophosphorylation and may involve binding of SH2 domain proteins to autophosphorylation sites (34). The second is independent of autophosphorylation and likely involves NGF binding simultaneously to two gp75 molecules, perhaps analogous to the ligand-receptor complex for tumor necrosis factor (35). Such bridges between gp75 molecules would lead to clustering and immobilization of the gp75TrkA complexes. We propose that gp75 stabilizes and promotes the ability of NGF to cross-link and immobilize TrkA. This immobilization may restrict TrkA to signaling complexes or sites within the membrane and thereby spatially restrict autophosphorylation and signal transduction.
Biochemistry, Vol. 37, No. 9, 1998 3179
FIGURE 1: Schematic of TrkA, indicating tyrosine phosphorylation sites: the SHC binding site at Y490, sites within the tyrosine kinase domain at Y670, Y674, and Y675, and the PLCγ1 site at Y785.
EXPERIMENTAL PROCEDURES Antibodies and Fragments. Rabbit anti-TrkA serum RTA (36) was generously supplied by Dr. Louis Reichardt (UCSF School of Medicine, San Francisco). IgG was purified from this antiserum on a Bio-Rad protein A column modified to increase the yield by using 100 mM Tris, pH 8.0, as the binding buffer, 10 mM Tris, pH 8.0, as the wash and 100 mM glycine, pH 3.0, as the eluting buffer. Fab fragments of these mAb’s were prepared as previously described (37, 38). A fluorescein Fab fragment of goat antirabbit IgG was obtained from Cappel (Durham, NC). Rabbit antiserum 203 against the Trk C-terminus (39) was used for immunoblotting TrkA. Recombinant anti-phosphotyrosinehorseradish peroxidase conjugate (RC20) was obtained from Transduction Laboratories (Lexington, KY) and used for immunoblotting. BaculoViruses. Recombinant baculoviruses for human wild-type gp75 and mutant human TrkA were prepared as described previously (32, 33, 40). The structure of TrkA is shown in Figure 1, indicating the major regions: extracellular, transmembrane, and cytoplasmic domains. Also indicated are the sites of tyrosine phosphorylation. Mutations used in this study are TrkA (K538N), a point mutation at the ATP binding site of the kinase domain, TrkA (Y490F), which does not contain the SHC phosphorylation site, TrkA (Y785F), which lacks a PLCγ1 site, TrkA (YY/490,785/FF), which lacks both of these sites, and TrkA (YYY/670,674,675/FFF), which is a triple-point mutation in the regulatory region of the kinase domain, which is kinase-deficient (16, 17) like TrkA (K538N). Cell Lines. Rat pheochromocytoma PC12 cells were maintained in DMEM supplemented by 10% heat-inactivated horse serum, 5% heat-inactivated fetal bovine serum, and 100 µg/mL gentamicin at 37 °C under 5% CO2 humidified atmosphere. Sf9 insect cells were maintained in TMN-FH medium from JRH Biosciences (Lenexa, KS) supplemented with 10% heat-inactivated fetal bovine serum and 100 mg/ mL gentamicin at 28-29 °C. BaculoVirus Expression in Sf9 Cells. To express a single NGF receptor, recombinant baculovirus was added to Sf9 cells (2 × 106 cells in a 25-cm2 flask) at a multiplicity of infection (moi) of 1 for the gp75 virus or a moi of ∼10 for TrkA. For coexpression experiments, TrkA baculovirus was added to Sf9 cells at a moi of ∼10 followed by immediate addition of gp75 baculovirus at moi of 1. These conditions resulted in infected cells displaying high-affinity binding of NGF and a ratio of gp75/TrkA proteins similar to that observed for Trk-PC12 cells which are highly responsive to NGF (32, 39).
3180 Biochemistry, Vol. 37, No. 9, 1998
Wolf et al.
Table 1: Diffusion Data for TrkA on PC12 and Sf9 Cellsa NGF receptors
cell
R (%)
SD (% R)
D ×109 s/cm2
SD ×109 s/cm2
n
gp75 + TrkA gp75 + TrkA + NGF Gp75 + K538N Gp75 + K538N + NGF Gp75 + TrkA Gp75 + TrkA + NGF K538N K538N + NGF YYY/670,674,675/FFF YYY/670,674,675/FFF + NGF TrkA TrkA + NGF YY/490,785/FF YY/490,785/FF + NGF Y490F Y490F + NGF Y785F Y785F + NGF
PC12 PC12 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9 Sf9
28 ( 1 21 ( 1 43 ( 2 25 ( 1 43 ( 2 29 ( 0.7 34 ( 1 36 ( 1 32 ( 1 35 ( 2 46 ( 2 23 ( 1 30 ( 1 33 ( 1 47 ( 2 48 ( 2 38 ( 1 33 ( 1
10 9 18 7 22 10 16 14 18 20 21 7 16 15 17 19 16 13
3.64 ( 0.23 4.11 ( 0.18 3.81 ( 0.36 1.64 ( 0.38 2.31 ( 0.25 2.17 ( 0.29 2.72 ( 0.47 2.27 ( 0.25 2.89 ( 0.25 2.62 ( 0.24 2.64 ( 0.21 1.73 ( 0.11 2.59 ( 0.23 3.04 ( 0.39 1.95 ( 0.14 2.66 ( 0.20 2.03 ( 0.16 2.59 ( 0.19
2.26 1.87 5.42 1.19 3.99 3.44 5.90 2.83 2.94 2.55 3.16 1.27 2.74 3.74 1.54 2.09 1.39 1.98
99 106 206 190 256 130 162 133 112 116 233 148 148 98 120 107 125 114
a
D and R given as mean ( standard error of the mean. SD, standard deviation; n, number of single bleach measurements.
Measurement of TrkA Autophosphorylation. Cells (6 × 106/sample) were treated for 10 min with 4 nM NGF in complete medium. The cells were extracted at 4 °C with 0.5% (v/v) NP-40, 140 mM NaC1, 10 mM Tris, 10 mM NaF, 5 mM Na2EDTA, 100 kallikrein IU/mL of aprotinin, 10 mM Na2VO4, and 1 mM phenylmethanesulfonyl fluoride, pH 7.5, and then clarified by centrifugation at 76000g for 1 h. TrkA was immunoprecipitated with 203 anti-Trk C-terminus rabbit serum (1:500) (39) and applied to a 10% SDS polyacrylamide gel. After electrophoresis, the proteins were electrotransferred to an Immobilon-P membrane (Millipore, Bedford, MA), followed by blocking for 1 h at 37 °C with 1% BSA (w/v), 0.1% Tween-20 (v/v), 137 mM NaCl, 20 mM Trizma, pH 7.6. The membrane was then incubated for 1 h at 37 °C with an RC20 anti-phosphotyrosine antibody-horseradish peroxidase conjugate (0.2 mg/mL) (Transduction Laboratories). Reactive bands were visualized with a chemiluminescent substrate (41). The membrane was then stripped by a 45-min incubation at 70 °C in 2% SDS (w/v), 100 mM β-mercaptoethanol, 62.5 mM Trizma, pH 6.8. The membrane was then probed with 203 anti-Trk antiserum, washed, and incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG secondary antibody (1:1000) (Amersham, Arlington Heights, IL) and finally with chemiluminescent substrate. Labeling of Cells for FRAP Measurements. For FRAP measurements, PC12 cells (5 × 105) were incubated for 30 min at room temperature with 50 µL of an anti-TrkA Fab fragment of rabbit antibody RTA (0.1 mg/mL in Dulbecco’s Minimum Essential Medium (DMEM) supplemented with 1% fetal bovine serum and 20 mM HEPES, pH 7.4). The samples were centrifuged, and the cells were washed twice with 200 µL of medium. The cells then were incubated for 30 min with 25 µg/mL fluorescein Fab fragment of antirabbit IgG. The cells were then pelleted through a cushion of DMEM with 5% fetal bovine serum. Since serum lacks detectable NGF, the serum present in these washes does not obscure the effects of added NGF (42). FRAP measurements on Sf9 cells were carried out ∼60 h postinfection. Cells (2.5 × 105) were incubated for 30 min at room temperature with 25 µL of a Fab fragment of RTA antibody in Sf9 growth medium (0.1 mg/mL). The samples
were washed by centrifugation (53g) for 3 min. The cells were suspended in 25 µg/mL fluorescein Fab fragment of anti-rabbit IgG and incubated for 20 min at room temperature. The cells were washed by centrifugation and then were pelleted through a cushion of TMN-FH medium with 10% fetal bovine serum. FRAP measurements were performed as previously described (37, 38). For measurements made in the presence of NGF, the cells were resuspended in medium containing 4 nM of NGF immediately prior to measurements. For both PC12 and Sf9 cells, all measurements were made at room temperature within 30 min of labeling, during which time no detectable internalization was observed. FRAP Measurements. The specific designs of our FRAP instrument and data analysis algorithm have been described in detail (43). All FRAP measurements were made as described previously (37, 38) at room temperature using a Leitz 63× na 1.4 planapochromat objective and the 488-nm line of a Lexel 95-2 argon laser. At the object plane of the microscope, the laser beam in this system has the form
I(x,y) ) Io exp (-2(x2 + y2)/w2) where Io is the intensity at the center, x and y are the Cartesian coordinates in the plane of the object, and w is the beam radius, 0.9 µm. The monitoring intensity was 0.13 µW, and the bleaching intensity was 1.3 mW for 25 ms. These conditions were chosen so that there would be no significant bleaching due to the monitoring beam. Samples were discarded if solution background intensities exceeded 10%. Data were fitted to the diffusion theory of Axelrod et al. (44) by a modification of the nonlinear least-squares procedure of Bevington (43, 45, 46). FRAP data are presented as averages ((SEM) of n single bleach measurements made on n separate cells. Additionally, the standard deviations are given in Table 1 as a measure of the widths of the distributions. Typical FRAP recovery curves are shown in Figure 2 for TrkA expressed in Sf9 cells in the absence and presence of NGF. If F(t < 0) is the prebleach fluorescence intensity, F(0) the fluorescence intensity immediately following the bleach, and F(∞) the fluorescence intensity after recovery is
NGF Receptor TrkA Membrane Diffusion
Biochemistry, Vol. 37, No. 9, 1998 3181
FIGURE 2: Typical FRAP recovery curves showing the diffusion of TrkA expressed in Sf9 insect cells in the absence or presence of NGF. Data have been normalized by dividing by the average prebleach intensity so that F(t
